Quality Assurance

The system seeks to detect errors, bias, and risk early; assign responsibilities; document decisions; and reduce the chance that a design problem reaches fieldwork or that a data problem reaches the report. Quality assurance does not mean research has no limitations; it means limitations are identified, managed, and disclosed.

• Study lead: overall coherence, questions, and methodology.
• Field lead: training, allocation, monitoring, and safety.
• Data lead: structure, cleaning, access, and versioning.
• Analyst: rules, code, tables, and interpretation.
• Reviewer: relatively independent review of figures, narrative, and limitations.
• Authorized approver: final approval or request for revision.

• Review scope, questions, and deliverable matrix.
• Assess feasibility, access, and risk.
• Review sample, size, and assumptions.
• Review instrument linguistically, methodologically, and ethically.
• Test programming, routing, and values.
• Pilot and document changes.
• Approve field, quality, and security plans.

• Track progress against sample or quotas.
• Review geographic and time distribution.
• Check impossible values, duplication, and patterns.
• Review duration and routing.
• Verify or recontact under the consent and quality plan.
• Observe researchers or review authorized recordings.
• Suspend work where a material failure appears.

A record is not deleted and a researcher is not accused on the basis of one signal. Indicators are combined, the case is reviewed, and the conclusion documented. Action may include retraining, verification, recollection, record exclusion, researcher suspension, or plan revision. The effect on findings and schedule is assessed.

• Clear variable and value templates.
• Log of data changes and reasons.
• Separation of raw and processed versions.
• Encryption or protection of sensitive files.
• Limited permissions and institutional accounts.
• Backups and restoration testing.
• Identification of the version used for each output.

• Reproduce tables from documented code or rules.
• Review derivations, weights, and denominators.
• Match totals, percentages, and missing values.
• Compare versions and data sources.
• Review small cells and multiple testing.
• Challenge alternative interpretations and causal claims.

• Match every figure to an approved table or output.
• Ensure consistency across narrative, tables, charts, and summary.
• Include fieldwork dates, scope, and methods.
• Explain limitations and uncertainty.
• Review Arabic and English without changing meaning.
• Review privacy, rights, and citation.
• Manage version number and approval date.

Material deviation from protocol is logged with its reason, authorization, effect, corrective action, and disclosure requirements. Issue logs are used to improve instruments, training, and systems in later projects.

• Response and completion rates.
• Error, recollection, and exclusion rates.
• Coverage against plan.
• Interviewer and interview consistency.
• Table or version errors.
• Time to resolve review comments.
• Participant complaints or privacy incidents.
• Completion of corrective action.